Enhancing engagement and self-efficacy through AI-supported Chinese language instruction: a mixed-methods intervention study

Abstract

This study examined the impact of AI-supported Chinese language instruction on engagement and self-efficacy among Malaysian university students. A quasi-experimental intervention with two instructional groups, an AI-supported teaching group and a traditional teaching group, was employed using an explanatory sequential mixed-methods design. The three-month intervention involved 100 undergraduate students. Measures of engagement and self-efficacy were collected before and after the intervention to examine changes within and between groups. Mixed-design ANOVA revealed significant Group × Time interaction effects for both engagement (partial η2 = 0.165) and self-efficacy (partial η2 = 0.116), indicating significantly greater improvements in the AI-supported group than in the traditional group. Results showed that students in the AI-supported group experienced significant gains in engagement and self-efficacy, whereas the traditional group showed little change. Following the quantitative phase, semi-structured interviews were conducted to explain the quantitative findings. Qualitative analyses helped explain these findings, indicating that AI-supported activities fostered active engagement, enhanced cognitive processing, promoted positive emotional experiences, and increased confidence in Chinese language knowledge and production. The study highlights the potential of AI tools to enhance motivation and psychological outcomes in language learning and provides practical recommendations for implementing AI in the classroom through effective teacher support and task design.